grep wont search multiple patterns sepperated by a newline

[The Cog]

I thought that wrapping a string in quotes was sufficient to protect the meta-characters in it from being expanded by the shell? why is the '$' necessary? (or does the '$' ensure that '\n' is expanded by the shell??)

Exactly that. "man bash" is really hard to read unless you know what you are looking for. Try "man bash" and search for QUOTING (yes, capitals). There is this sentence:

Words of the form $'string' are treated specially. The word expands to string, with backslash-escaped characters replaced as specified by the ANSI C standard.

[Holger_Gehrke]

From 'man bash', section "QUOTING":

... Words of the form $'string' are treated specially. The word expands to string, with backslash-escaped characters replaced as specified by the ANSI C standard. ....

So, yes, the '$' makes the bash expand '\n' to a newline-character. (Don't know how often I've read that page over the years and never noticed this construct ...)

I don't have backslashes before the newlines in my example because those would make grep give me an error:"grep: Angehängter Backslash (»\«)" ('appended backslash'). Without the backslash your example would search for the patterns 'owl' and '*****-cat' in the file 'theowlandthepussycat'. If those aren't in there, you'd get no output.

Holger

[TheFu]

This actually works! it was the '$' that I was missing. I thought that wrapping a string in quotes was sufficient to protect the meta-characters in it from being expanded by the shell? why is the '$' necessary? (or does the '$' ensure that '\n' is expanded by the shell??)

Perhaps I missed it, but this is a bash-specific thing. Other shells would do it different. Bash expansion (regex/globbing) is mostly intuitive, until it isn't.

If it were me, I'd use egrep (the same as grep -E) which is for 'enhanced regex grep'. 2 fewer characters to type, 'eg{tab}' does the command completion fast.

       -E, --extended-regexp
              Interpret PATTERNS as extended regular  expressions  (EREs,  see
              below).

       -P, --perl-regexp
              Interpret  PATTERNS  as  Perl-compatible   regular   expressions
              (PCREs).   This option is experimental when combined with the -z
              (--null-data) option, and grep  -P  may  warn  of  unimplemented
              features.

This basically results in nearly python/perl regex handling, so there are fewer surprises. I know perl, so those regex are sometime needed, -P option, but that is seldom, since whenever a bash script gets over 1 page in length, I switch to perl for all the added capabilities like multi-line regex codes with comments. Perl hasn't been write-only in a very long time.

If multiple regex are needed, consider using multiple -e options, one for each.

[jcdenton1995]

From 'man bash', section "QUOTING":

So, yes, the '$' makes the bash expand '\n' to a newline-character. (Don't know how often I've read that page over the years and never noticed this construct ...)

I don't have backslashes before the newlines in my example because those would make grep give me an error:"grep: Angehängter Backslash (»\«)" ('appended backslash'). Without the backslash your example would search for the patterns 'owl' and '*****-cat' in the file 'theowlandthepussycat'. If those aren't in there, you'd get no output.

Holger

Righhht now I've got it, I never realised you can just hit enter after typing the first pattern to move to a newline in the terminal (naturally I assumed this would just attempt to run an incomplete command), so is that specifically a function of grep, that it 'knows' when your giving pattern arguments, if you hit enter it should simply move to a new line?

Anyway we've established that one can search for multiple patterns in the following three ways...

grep $'[PATTERN1]\n[PATTERN2]\n[PATTERN3]' [INPUTFILE]


grep '[PATTERN1]
[PATTERN2]
[PATTERN3]' [INPUTFILE]


grep -E '[PATTERN1]|[PATTERN2]|[PATTERN3]' [INPUTFILE]

So that pretty much clears it up, thank you all for your help!

[TheFu]

This should work too:

grep -E -e [PATTERN1] -e [PATTERN2] -e [PATTERN3]  [INPUTFILE]

or if you want them on separate lines:

grep -E -e [PATTERN1] \
-e [PATTERN2] \
-e [PATTERN3]  \
[INPUTFILE]

Just ensure there isn't any whitespace after the normal line continuation character \.

[Holger_Gehrke]

Righhht now I've got it, I never realised you can just hit enter after typing the first pattern to move to a newline in the terminal (naturally I assumed this would just attempt to run an incomplete command), so is that specifically a function of grep, that it 'knows' when your giving pattern arguments, if you hit enter it should simply move to a new line?

It's not a function of grep, grep isn't running yet. It's the shell (or probably the readline library that the shell uses for text entry) that recognizes the odd number of quotes as a sign that there's more to come.

Holger